Project Summary/Abstract With the global rise in obesity-related diseases and increasing recognition of the role of fat in diseases such as fatty liver, there is growing need for a noninvasive chemical assay of in vivo lipid depots. But the rich chemistry of these molecules is often ignored in MRI, which typically treats fat as a homogeneous proton species because differences in chemical shift and relaxation are difficult to distinguish in a clinical setting. Innovation: This proposal will be the first to use nonuniform echo spacing to generate J-coupling based contrast between chemically different fats. While fat spins often give monotonically decaying signal under a uniformly spaced spin echo train, J-coupling causes nonuniformly spaced spin echo trains to generate complex fingerprints, which distinguish and even quantify fats with otherwise similar MR properties. Significance: The immediate clinical impact will be to demonstrate a novel J-coupling based contrast mechanism in vivo which will contribute to a multiparametric MRI protocol to distinguish the 25% of the population with benign steatosis from the 2-3% with nonalcoholic steatohepatitis (NASH). The broader significance of this work is that this novel contrast mechanism could revolutionize the study and clinical management of many diseases associated with bioactive lipid subtypes. Approach: The aims are designed to test the hypothesis that nonuniform echo trains (1) can create contrast between different fats, which (2) will contribute to a multiparametric noninvasive protocol to distinguish benign steatosis and NASH. Aim 1 will identify the echo spacing best suited for distinguishing benign steatosis and NASH, first using density matrix simulations based on high resolution spectra of liver fat from animals with each disease, and then using experiments on whole rat livers. In Aim 2, the sequence developed in Aim 1 will be tested in humans with biopsy proven benign steatosis or NASH, along with a suite of standard methods of liver and fat characterization. This metric, and also metrics derived from conventional images, will be tested for the ability to separate these patient groups. In addition, machine learning algorithms will be used to find a metric that combines multiecho intensities with standard MR metrics. This proposal will prove the usefulness of nonuniform echo trains as a new contrast mechanism and yield a multiparametric MRI protocol to distinguish simple steatosis and NASH. Moreover, this general approach to distinguish fats could ultimately find a broad range of biomedical applications, such as imaging lipid based therapeutics, distinguishing the 4 fatty acids that make up 90% of the adipose stores, or detecting other lipotoxic subtypes. The results of this work will justify a larger scale study to properly validate the sensitivity to different fat species in humans.